CN114689015A - Method for improving elevation precision of optical satellite stereoscopic image DSM - Google Patents

Method for improving elevation precision of optical satellite stereoscopic image DSM Download PDF

Info

Publication number
CN114689015A
CN114689015A CN202111433232.2A CN202111433232A CN114689015A CN 114689015 A CN114689015 A CN 114689015A CN 202111433232 A CN202111433232 A CN 202111433232A CN 114689015 A CN114689015 A CN 114689015A
Authority
CN
China
Prior art keywords
dsm
elevation
transformation
photons
registration
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111433232.2A
Other languages
Chinese (zh)
Other versions
CN114689015B (en
Inventor
叶江
强宇轩
张铃
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chengdu Univeristy of Technology
Original Assignee
Chengdu Univeristy of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chengdu Univeristy of Technology filed Critical Chengdu Univeristy of Technology
Priority to CN202111433232.2A priority Critical patent/CN114689015B/en
Publication of CN114689015A publication Critical patent/CN114689015A/en
Application granted granted Critical
Publication of CN114689015B publication Critical patent/CN114689015B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C5/00—Measuring height; Measuring distances transverse to line of sight; Levelling between separated points; Surveyors' levels
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C11/00—Photogrammetry or videogrammetry, e.g. stereogrammetry; Photographic surveying
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/88—Lidar systems specially adapted for specific applications
    • G01S17/89—Lidar systems specially adapted for specific applications for mapping or imaging
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/88—Lidar systems specially adapted for specific applications
    • G01S17/89—Lidar systems specially adapted for specific applications for mapping or imaging
    • G01S17/894—3D imaging with simultaneous measurement of time-of-flight at a 2D array of receiver pixels, e.g. time-of-flight cameras or flash lidar
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G06T17/05—Geographic models

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Software Systems (AREA)
  • Geometry (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Electromagnetism (AREA)
  • Computer Graphics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Image Processing (AREA)

Abstract

The invention relates to an optical satellite stereoscopic image technology. Aiming at the problems that the lack of ground actual measurement control points affects the elevation precision of the optical satellite stereoscopic image generated DSM and further limits the application of high-resolution DSM in the fields of surveying and mapping, hydrology, geology and the like, the invention provides a method for improving the elevation precision of the optical satellite stereoscopic image DSM, which utilizes the satellite stereoscopic image to establish the DSM and obtains satellite laser height measurement data of the same area; and after the space coordinate transformation is carried out on the two, the registration is carried out, and the registration state with the minimum standard deviation of the elevation difference of the two is taken as the registration result, so that the elevation precision of the DSM is improved. By utilizing the characteristics that the ICESat-2 satellite height measurement data is high in elevation precision, densely distributed along straight tracks on the ground and large in track distance, a plurality of ICESat-2 tracks are combined together to approximate to a DSM with excellent elevation precision and high resolution along the direction of a laser track, and therefore the DSM registration technology is utilized to improve the geometric precision of the DSM generated by the optical satellite stereoscopic image.

Description

Method for improving elevation precision of optical satellite stereoscopic image DSM
Technical Field
The invention relates to an optical satellite stereoscopic image technology, in particular to an optical satellite stereoscopic image elevation precision technology.
Background
With the continuous development of remote sensing and photogrammetry technologies, high-resolution stereoscopic images become the main way to acquire high-resolution dsm (digital Surface model) in a large range. The method benefits from the progress of two technologies, namely, the sub-meter-level three-dimensional remote sensing image imaging technology and the sensor types tend to be diversified; and secondly, the high-resolution remote sensing image processing technology greatly promotes the acquisition efficiency of the DSM. Meanwhile, with the progress of research, acquisition of higher-precision DSMs is required in scientific research fields such as mapping, geology, hydrology and the like. According to the working principle of the satellite linear array push-broom type imaging sensor, in the satellite design, the main factors influencing the positioning precision of the satellite image comprise the orbit determination precision, the attitude determination precision, the sensor precision, the precision of each time information and the like. The conventional techniques for improving the positioning accuracy of satellite images in photogrammetry include: a. geometric calibration of the sensor; b. calculating the adjustment of the area network; in the former, the error characteristics of equipment systems such as an imaging sensor, a star sensor and the like need to be analyzed for a long time to ensure the consistency of positioning accuracy of remote sensing images in the global scope; the latter needs to use extra information to correct errors in relevant parameters of a target image, and has the difficulty that ground actual measurement control points cannot be always obtained, and in areas with inconvenient traffic, such as plateau mountain areas, enough control points cannot be generally obtained, so that the use of the area network adjustment technology is limited by regions; c. the error of the DSM caused by the imaging condition is compensated by using a three-dimensional point cloud registration method, and the geometric accuracy of the DSM can also be improved. Due to the development of a high-precision attitude determination system, the satellite image plane positioning precision can obtain an ideal result under the condition of no control point, but the elevation precision improving effect is weaker than the plane precision, such as: under the condition of no control point, the Worldview-2 image has the plane positioning precision of 2.3m and the elevation precision of 5 m.
In order to obtain Elevation data in a global range, quantify the influence of glacier ablation on sea level rise, estimate global biomass and monitor terrain change, a second Ice, Cloud and land Elevation Satellite ICESat-2(Ice, Cloud and land Elevation Satellite-2) is transmitted by the United states aerospace administration (NASA) in 2018, 9 and 15, and an Advanced Terrain Laser Altimeter System (ATLAS) carried on the NASA adopts a micro-pulse photon counting Laser radar technology, and the measurement precision reaches 10 cm. The distance between adjacent laser points along the track direction is 70cm, the diameter of a laser footprint is about 17m, the final diameter can reach 20m along with energy consumption, 6 beams of laser can be emitted simultaneously, every two beams of strong laser and weak laser form one group, the distance between the two adjacent groups is 3.3km in the direction perpendicular to the ground track, and the distance between the laser in each group is 90 m. The mode of combining the strong and weak laser beams is convenient for the ATLAS to acquire enough returned photons with the earth surface with high reflectivity at low reflectivity and save energy, and 3 groups of strong and weak laser beams form cross measurement, so that the detection capability of the ATLAS on the slope and the elevation change of the earth surface can be effectively improved. The ATLAS Data product is divided into 5 levels, all of which are available at the National Snow & Ice Data Center (NSIDC). The ATL08 is a 3A-level product, records the elevation information of land and vegetation canopies, also comprises longitude and latitude coordinates of each laser point and related quality evaluation parameters, the spatial resolution along the track direction is 100m, the time resolution is 91 days, and the data covers the whole world. The ATL08 is obtained by processing the ATL03 data by a DRAGANN filtering algorithm. ATL03 is L2 grade global geolocation photon data with a spatial resolution of 0.7m, the recorded information contains the height, latitude, longitude and photon flight time of each photon on the WGS84 ellipsoid, the ATL03 product is a bridge between low-grade, instrumental (ATL01/02) and high-grade, surface-specific, scientific research-centric (ATL06 and above), providing all the information needed for higher level data products. In addition, the ATL03 data classified each photon as a signal photon or a background photon, which in turn was identified as five classes of land ice, sea, land and inland waters, and provided confidence estimates for these classes. ATL03 data applies various geophysical parameter corrections, optimizing elevation for each photon, allowing the user to edit these geophysical parameters as desired.
Since ICESat-2 data release, the method plays an important role in the fields of offshore water depth measurement, glacier ablation and height change, forest fire area detection, forest biomass estimation, forest canopy height measurement and the like. The ICESat-2 precise Pointing Positioning (PPD) technology enables the ICESat-2 laser point plane geographic positioning Precision to be better than 6.5 m. The Elevation location of the expression of the ATL08 product is between the 3DEP (3-D Elevation program) DEM (Digital Elevation Model, DEM for short) and the actual surface Elevation. Experiments show that the selection of proper parameters for screening can obtain high-precision control points from ATL 08. However, for ICESat-2 laser elevation data, since no laser spot footprint image is recorded, the application in area grid leveling is limited. In addition, if the DEM is constructed by adopting the discrete point clouds through a spatial interpolation method, the result is easily influenced by the spatial distribution state of the discrete point clouds, and the method is not suitable for ICESat-2 data. The plane positioning accuracy of the ATLAS is better than 6.5m, the elevation accuracy is 0.2m in plain areas and 2.0m in mountain areas, and the ATLAS is better than the elevation accuracy when the remote sensing images are directly positioned on the ground. It can be seen that it is a feasible idea to improve the elevation accuracy of optical stereo images by using the ICESat-2 satellite laser height measurement data, but the two data must be accurately matched.
Disclosure of Invention
The invention provides a method for improving the elevation precision of an optical satellite stereoscopic image, aiming at the problems that the lack of ground actual measurement control points affects the elevation precision of an optical satellite stereoscopic image to generate DSM and further limits the application of high-resolution DSM in the fields of surveying and mapping, hydrology, geology and the like, and provides a method for improving the elevation precision of an optical satellite stereoscopic image.
The invention adopts the technical scheme that the method for improving the elevation precision of the optical satellite stereoscopic image DSM comprises the following steps: creating a DSM (digital surface model) by using a satellite stereoscopic image to acquire satellite laser height measurement data of the same area; and after the space coordinate transformation is carried out on the two, the registration is carried out, and the registration state with the minimum standard deviation of the elevation difference of the two is taken as the registration result, so that the elevation precision of the DSM is improved.
Specifically, when the satellite stereogram is used for creating the DSM, the image geometric model adopts an RFM model, the output resolution is not more than 2 times of the resolution of the image, and the image geometric model is stored in a regular grid form.
Specifically, the method comprises the following steps: the method comprises the steps of obtaining satellite laser height measurement data of the same area, preprocessing the satellite laser height measurement data to obtain high-quality laser height measurement data, and then performing space coordinate transformation; the pretreatment comprises the following steps:
weak energy laser beam data in the satellite laser height measurement data are removed, and medium and high-confidence photons when the earth surface coverage type is the land are selected from the rest data; the medium and high confidence photons are linearly gathered and distributed along the laser track, the medium and high confidence photons are distributed on a line segment perpendicular to the direction of the laser track, and the average elevation value of the photons in the same line segment is taken as the final elevation of the line segment; and removing photons with abnormal elevation.
Specifically, the method for eliminating the photons with abnormal elevation comprises the following steps:
and setting a threshold value of the height difference between the photons and the DSM, and removing the photons exceeding the threshold value as the photons with abnormal height.
Further, after the space coordinate transformation is carried out on the two, the registration is carried out, and the registration state with the minimum standard deviation of the elevation difference between the two is taken as the registration result, so that the elevation precision of the DSM is improved, and the method comprises the following steps:
after space coordinate transformation is carried out on DSM and high-quality laser height measurement data, the maximum translation rotation amount is set, the corresponding point relation is established, the high-quality laser height measurement data is used as source data, DSM is used as reference data, and space rotation transformation is carried out around an X, Y, Z axis; translation transformation is carried out in an X-O-Y plane, and the standard deviation of the height difference of the corresponding point under different transformation states is calculated;
selecting the rotation translation amount with the minimum standard deviation of the height difference as a transformation parameter;
and taking the DSM as source data, and performing inverse transformation according to the transformation parameters so as to improve the elevation precision of the DSM.
Further, the method also comprises the following steps: and under different transformation states, rejecting photons with abnormal elevation in the high-quality laser height measurement data under the state, and recalculating the standard deviation of the height difference of the corresponding point under the transformation state.
Further, the calculation formula of the standard deviation corresponding to the height difference of the points under different transformation states is as follows:
ΔH＝{Δh(i,j)/Δh(i,j)＝HATL(i,j)-HDSM(i,j),i∈1,2…n,j∈1,2…m}；
wherein HATL(i,j)、HDSM(i, j) respectively indicating the elevations of corresponding points formed by laser points of high-quality laser height measurement data and DSMs, wherein delta H (i, j) is the elevation difference of the corresponding points after each transformation, and delta H is an elevation difference set of the corresponding points in all transformation states; i is the number of corresponding points, and j is the number of times of executing spatial transformation on the high-quality laser height measurement data;
Figure BDA0003380834250000031
wherein,. DELTA.hsd(j) Standard deviation of elevation difference, delta H, of corresponding point under jth spatial transformation for high-quality laser altimetry dataSDAnd (4) setting standard deviation of elevation difference of corresponding points in all transformation states.
Further, the calculation formula of the transformation parameter is as follows:
Figure BDA0003380834250000032
wherein T is a translation matrix; r is a rotation matrix; alpha, beta, gamma, delta x, delta y and delta z are respectively the rotation translation amounts of the x axis, the y axis and the z axis; the translation parameter Δ z in the z-axis direction is an elevation difference average value when the standard deviation of the elevation difference is minimum.
Still further, the calculation formula for rejecting the photons with elevation abnormality is as follows:
Figure BDA0003380834250000041
wherein Hu_ATL(i, j) high-quality laser height measurement data with abnormal elevation removed in the jth transformation state, Hut_ICESAT2High-quality laser height measurement data sets with the elevation outliers removed in all the transformation states are acquired;
then there are:
ΔHu＝{Δhu(i,j)/Δhu(i,j)＝Hu_ATL(i,j)-HDSM(i,j),i∈1,2…n,j∈1,2…m}；
wherein, Δ HuFor the recalculated set of elevation differences, Δ h, for all transform statesu(i, j) is the recomputed height difference for the j-th transformation.
Still further, the calculation formula of the standard deviation of the minimum height difference is as follows:
Figure BDA0003380834250000042
wherein,. DELTA.husd(j) The standard deviation of the height difference recalculated under the same transformation state; Δ husd_minThe standard deviation of the smallest elevation difference for all the transition states.
The method has the advantages that according to the scheme, multiple ICESat-2 tracks are combined together to approximate a DSM with superior elevation accuracy and high resolution along the direction of a laser track by utilizing the characteristics that the ICESat-2 satellite height measurement data are high in elevation accuracy, are densely distributed along straight tracks on the ground and have large track distance, a DSM registration strategy (BOTS) Based on Terrain Similarity is provided, an objective function is established between the two DSMs, the objective function is minimized by moving and rotating the DSM, and the two DSMs are aligned in a three-dimensional space. Therefore, under the condition of no ground actual measurement control point, ICESat-2 laser height measurement data is assisted to improve the elevation precision of DSM generated by the optical stereo image of the remote sensing satellite.
Drawings
FIG. 1 is a distribution diagram of experimental regions in an example of the present invention.
FIG. 2 shows the terrain and ATL03 data and the range of stereo images in each experimental area according to an embodiment of the present invention.
Fig. 3 is a flow chart of DSM registration based on terrain similarity in an embodiment of the present invention.
FIG. 4 is a flow chart illustrating the pre-processing of ATL03 data according to an embodiment of the present invention.
FIG. 5 is a cloud diagram of ATL03 photons in the same trace according to an embodiment of the present invention.
Fig. 6 is a schematic diagram of DSM registration based on spatial coordinate transformation in an embodiment of the present invention.
Fig. 7 is a flowchart of a DSM registration method based on terrain similarity in an embodiment of the present invention.
FIG. 8 is a DSM extracted from 5 pairs of voxels using an SGM algorithm in an embodiment of the present invention; a. b, c, d and e are respectively: WV2_ DSM, ZY3_ DSM _ N32E91, GF7_ DSM, ZY3_ DSM _ N39E115, SV _ DSM.
FIG. 9 is a DSM center plane translation after registration in an embodiment of the invention; the left graph shows the amount of movement of each DSM center in the east-west direction, and the right graph shows the amount of movement of each DSM center in the north-south direction.
FIG. 10 is a DSM elevation residual distribution in an embodiment of the invention; a. b, c, d and e are quartiles and mean values of elevation residual absolute values of DSMs before and after ICP and BOTS algorithm registration.
FIG. 11 is a DSM elevation accuracy evaluation in an embodiment of the invention.
FIG. 12 is a DSM elevation residual spatial distribution after BOTS algorithm registration in an embodiment of the present invention; a. b, c, d and E respectively represent the spatial distribution of elevation residual absolute values after WV2_ DSM, ZY3_ DSM _ N32E91, GF7_ DSM, ZY3_ DSM _ N39E115 and SV _ DSM are registered by BOTS algorithm; the magnitude of the ATL03DSM elevation residual checkpoint symbols is taken to be the absolute magnitude of the elevation residuals, i.e., a larger symbol indicates a larger elevation residual.
Detailed Description
The technical solution of the present invention is described in detail below with reference to the embodiments and the accompanying drawings.
The invention relates to a method for improving the elevation precision of an optical satellite stereoscopic image, which comprises the following steps: creating a DSM (digital surface model) by using a satellite stereoscopic image to acquire satellite laser height measurement data of the same area; and after the space coordinate transformation is carried out on the two, the registration is carried out, and the registration state with the minimum standard deviation of the elevation difference of the two is taken as the registration result, so that the elevation precision of the DSM is improved. By utilizing the characteristics that the ICESat-2 satellite height measurement data is high in elevation accuracy, densely distributed along straight tracks on the ground and large in track distance, a plurality of ICESat-2 tracks are combined together to approximate a DSM with excellent elevation accuracy and high resolution along the direction of a laser track, a DSM registration strategy (BOTS) Based on Terrain Similarity is provided, an objective function is established between the two DSMs, the objective function is minimized by moving and rotating the DSMs, and the two DSMs are aligned in a three-dimensional space. Therefore, under the condition of no ground actual measurement control point, ICESat-2 laser height measurement data is assisted to improve the elevation precision of DSM generated by the optical stereo image of the remote sensing satellite.
Examples
In order to further fully illustrate the feasibility of the invention, the example selects different topographic conditions, adopts Worldview-2, SV-1, GF-7 and ZY-3 stereo images to verify the feasibility and effectiveness of the method, and compares the method with the most common Iterative Closest Point (ICP) algorithm at present.
To fully illustrate the effectiveness of the present invention, areas located at different locations, terrain, and types of surface coverage are selected as the areas of interest. As shown in FIG. 1, the area A is located at the south edge of the basin of the Chachida, the northeast slope of Kunlun mountain, the north latitude is 36 degrees 12 'to 36 degrees 20', and the east longitude is 95 degrees 42 'to 96 degrees 3'. Because the altitude is 2700m-4300m, the air is thin, the vegetation in the area is rare, most of the mountain bodies are bare in bedrock, and other areas have no vegetation except a few shrubs such as salix purpurea, reed and the like in the riverbed valley land, and belong to the plateau desert mountain area. The regional terrain is complex, the mountain is steep, the altitude drop is large, the regional terrain is a typical region for researching the positioning accuracy of the optical remote sensing image in the plateau mountain region to the ground, and the experimental result is representative. The research area B is positioned in the Tibet plateau, and the north latitude is 32 degrees 45 degrees to 33 degrees 21 degrees, and the east longitude is 91 degrees 40 degrees to 92 degrees 21 degrees. The elevation is between 4700m and 6100m, the plateau grassland is mainly planted in the area, no high vegetation exists, the mountain top is covered by the accumulated snow, the terrain in the area is complex, the mountain is steep, and the elevation drop is large. The research area C is located in the range of 27-27 degrees 20 'in north latitude and 75-75 degrees 57' in east longitude, and is located in the middle and downstream of the Yangtze river. The elevation is between 0m and 1400m, high and large vegetation covers the area, and the area belongs to mountain terrain, the mountain terrain is steep, and the elevation fall is large. The research area D is located in the range of North China plain, north latitude 39 degrees 7-39 degrees 40 degrees, east longitude 115 degrees 19-116 degrees 1'. The elevation is between 0m and 1400m, the crops are mainly planted in the area, the plain is mainly planted in the terrain, and the mountain land terrain is in the northwest area. Study area E was located in the range of india, 27 ° 10 'to 27 ° 20' north latitude, and 75 ° 46 'to 75 ° 56' east longitude. The elevation is 380m-750m, the crops are mainly planted in the area, the plain is mainly planted in the terrain, and the hills are in the east area. Table 1 describes the stereo images and the main parameters in each study area.
Table 1: stereo image and main parameters of research area
Figure BDA0003380834250000061
The ATL03 was chosen as the experimental reference point cloud data for the following reasons. a. Although both ATL08 and ATL03 provide terrain surface elevation data, ATL08 has a low spatial resolution along the track direction, a small number of laser spots, and cannot describe the terrain relief along the track direction in detail; b, the elevation recorded in ATL08 is the elevation of the terrain and the height of a vegetation canopy, the DSM is obtained through an optical remote sensing image and cannot penetrate through the terrain to obtain the elevation of the terrain, and the elevation recorded in ATL03 contains the elevation of the terrain on the ground, which is the same as the DSM; ATL03 has high spatial resolution along the track direction, can express topographic relief in detail and has huge point cloud number. The required ATL03 data was obtained by the described tool with time coverage ranging from 2018 to 2021. It should be noted that whatever data is used for registration is within the scope of the present invention. The topography and ATL03 data and the stereo image range in each experimental area are shown in fig. 2.
In the embodiment, DSM is extracted by taking a satellite panchromatic stereopair as basic data, ATL03 of the same area is obtained, the DSM registration strategy based on terrain similarity is utilized to perform space coordinate transformation on the satellite panchromatic stereopair and the satellite panchromatic stereopair, then registration is performed, and the registration state with the minimum standard deviation of the elevation difference of the satellite panchromatic stereopair and the satellite panchromatic stereopair is taken as a registration result, so that the elevation precision of the DSM is improved.
The specific DSM registration strategy is divided into the following 4 steps, as shown in fig. 3, (1) (2) data pre-processing. Eliminating ATL03 laser points with low confidence coefficient, and unifying spatial reference systems among multi-source data; (3) DSM registration based on terrain similarity. Calculating transformation parameters by taking DSM as reference data and ATL03 as source data; (4) an inverse DSM transform; taking the DSM as source data, performing inverse transformation according to the transformation parameters to register the DSM with the ATL03, and improving the geometric accuracy of the DSM.
The satellite three-dimensional image is used for creating the DSM, and the semi-global matching (SGM) is used as a dense matching method, so that the appearance of the elevation abnormal value is reduced to the greatest extent. According to experiments, the registration result when the dense data set is used as reference data has higher accuracy than that when the sparse data set is used as reference data, so that DSM extracted by a stereoscopic image is used as reference data, ATL03 laser height measurement data is used as data to be registered, and after the registration parameters are obtained, the conversion parameters are inversely applied to the DSM. The image geometric model adopts an RFM model, the output resolution does not exceed 2 times of the image resolution through optimizing a Rational Function Model (RFM), and DSM stores a regular grid form, which is convenient for processing calculation.
Although the ICESat-2 laser height measurement precision is high, the vertical precision of the laser point is unstable due to the influences of factors such as terrain, acquisition time, cloud layers and aerosol, and data points with large errors exist. Therefore, the ATL03 needs to be preprocessed before the registration starts, high-quality ATL03 laser elevation data are extracted, and the registration accuracy can be effectively improved. In this example, the ATL03 data preprocessing process specifically includes the following steps as shown in fig. 4:
first, ATL03 was screened for data:
and weak energy laser beam data in the satellite laser height measurement data are removed, and medium and high-confidence photons when the earth surface coverage type is the land are selected from the rest data. Research shows that the laser spot quality with a small h _ te _ uncertainties (total uncertainty of fitting elevation in 100m section) parameter value in the ATL08 data is high, while the h _ te _ uncertainties parameter value of the strong laser beam in the experimental area is generally lower than that of the weak laser beam, and the weak laser beam has large-range loss in the mountain area. Furthermore, as shown in fig. 5, all photons are separated into signal photons and background noise in the ATL03 dataset, and the confidence of each signal photon in land, sea ice, land ice and inland water is given by the signal _ conf _ ph parameter. When the value of signal _ conf _ ph is-2, -1, 0, 1, the photon is respectively indicated to belong to a tep (transmitter echo pulse) event photon, a photon irrelevant to a specific surface type, a noise photon, and a buffer photon, and when the value of signal _ conf _ ph is 2, 3, and 4, the photon is respectively indicated to belong to a low confidence level, a medium confidence level, and a high confidence level. In summary, weak energy laser beam data is first removed from ATL03 data, and then medium and high confidence ATL03 photons when the surface coverage type is land are selected from the remaining data as experimental data.
The value of the signal _ conf _ ph parameter has the following relationship with the type of ATL03 photons, wherein signal _ conf _ ph is-2, -1, 0 and 1 respectively represent that the photons belong to TEP event photons, photons unrelated to specific surface types, noise photons and buffer photons, signal _ conf _ ph is 2, 3 and 4 respectively represent that the photons belong to low confidence photons, and medium confidence photons and high confidence photons; here, photons of signal _ conf _ ph <2 are collectively referred to as noise photons; and the photons with medium and high confidence degree are linearly concentrated and distributed along the laser track.
Secondly, diluting the laser spot:
as shown in fig. 5, the medium and high confidence photons screened according to the parameters are linearly concentrated along the laser track. The medium and high confidence photons are distributed on a straight line segment which is approximately a straight line and is vertical to the laser track direction, the length of the line segment is about 0.5m, the distance between the line segments is 0.7m, the number of laser points on the same line segment is about 3-5, and the difference between the maximum elevation and the minimum elevation of the photons is 5m at most. The average elevation value of photons within the same line segment is taken as the final elevation value.
And finally, rejecting elevation abnormal data:
the ATL03 data quality is greatly improved through parameter screening, but at the moment, the ATL03 data contains multiple types of photons, such as: vegetation canopy photons, ice photons, terrain photons, and photons below bodies of water, and DSM expresses the elevation of terrain above the terrain, so photons with abnormal elevations need to be rejected before spatial transformation is performed. Photons, which are typically located below the feature, are removed as outliers. The core idea is that laser points exceeding the threshold are removed from ATL03 as abnormal data by setting the threshold of the height difference between the ATL03 laser point and DSM. For the threshold, the ATL03 laser point and DSM elevation difference for relatively smooth terrain and bare surface area were averaged. The strategy of iteratively culling ATL03 elevation anomaly data points will be adopted in this example to minimize the impact of ATL03 elevation anomaly data, as will be described in detail below.
As shown in fig. 6, when the sensor collects data, there is an unavoidable error between attitude and orbit measurement values, so that coordinate values of DSM data sets in the same area are inconsistent and cannot be overlapped in the same coordinate system. For the same data set, the data point acquisition conditions are the same, the system errors are equal, the coordinate inconsistency between the two DSM data sets can be equivalent to the translation rotation amount of one DSM data set in a coordinate system when the two DSM data sets are overlapped. Based on the idea, after the data processing is completed, the DSM registration based on terrain similarity is carried out on the data processing and the DSM registration, and the method comprises the following steps:
performing space coordinate transformation on DSM and high-quality ATL03 data, setting the maximum translation rotation amount, establishing a corresponding point relationship, taking high-quality ATL03 data as source data, taking DSM as reference data, and performing space rotation transformation around X, Y, Z axis; and translation transformation is carried out in an X-O-Y plane, and the standard deviation of the height difference of the corresponding point under different transformation states is calculated. Selecting the rotation translation amount with the minimum standard deviation of the height difference as a transformation parameter; and taking the DSM as source data, and performing inverse transformation according to the transformation parameters so as to improve the elevation precision of the DSM. As shown in particular in fig. 7.
First, the DSM and high quality laser altimetry data are spatially coordinate-transformed, and the transformed coordinates can be calculated according to the transformation matrix shown in equation (1), and since the DSM and ATL03 both represent absolute coordinates in WGS84 coordinates in this document, a scale factor is not considered in the coordinate transformation.
Figure BDA0003380834250000081
In the formula: t is a translation matrix; r is a rotation matrix; the alpha beta gamma is the rotation quantity of the x, y and z axes; Δ x Δ y Δ z are the translation amounts of the x, y, and z axes, respectively. The translation parameter Δ z in the z-axis direction is an elevation difference average value when the standard deviation of the elevation difference is minimum.
Secondly, the maximum translation rotation amount is set, so that the plane movement range is determined, unnecessary calculation amount is reduced, and the registration efficiency is improved. The maximum planar moving distance is determined by the direct ground positioning precision of the stereo remote sensing image and the planar precision of ICESat-2, and is different according to the image types.
And establishing a corresponding point relationship, specifically interpolating on DSM by using ATL03 to obtain a DSM elevation value, and forming a corresponding point with ATL 03. In order to weaken the correlation between the ATL03 data and a DEM (Digital Elevation Model, DEM for short), the relationship between the two data is expressed by the height difference of corresponding points.
The topography variations expressed by the two DSM data should be similar for topography within the same range, and when the two are aligned in the two-dimensional plane direction, the height differences at different plane coordinates should be equal, i.e., the standard deviation of the height differences should be minimal. Based on this, the DSM registration takes the standard deviation of the height difference as an objective function, and the calculation of the coordinate transformation parameters is divided into two steps, namely calculating the parameters of alpha, beta, gamma, delta x and delta y in the first step and calculating the translation parameter delta z of the z axis in the second step.
And (3) calculating the transformed coordinate values according to the formula (1) by using an ATL03 as a source data set and a DSM as a reference data set, performing spatial rotation transformation around X, Y and z axes, performing translation transformation in an X-O-Y plane, and calculating the standard deviation of the height difference of the corresponding point under different transformation states by using the formulas (2) and (3).
The calculation formula of the standard deviation of the height difference of the corresponding points under different transformation states is as follows:
ΔH＝{Δh(i,j)/Δh(i,j)＝HATL03(i,j)-HDSM(i,j),i∈1,2…n,j∈1,2…m}； (2)
wherein HATL03(i,j)、HDSM(i, j) are respectively the elevations of corresponding points formed by the laser points of the ATL03 and the DSM, wherein delta H (i, j) is the elevation difference of the corresponding points after each transformation, and delta H is the elevation difference set of the corresponding points in all the transformation states; i is the number of corresponding points, and j is the number of times of executing space transformation on the high-quality laser height measurement data;
Figure BDA0003380834250000091
wherein,. DELTA.hsd(j) Is the standard deviation, Δ H, of the elevation difference of the corresponding point under the j-th spatial transformation of ATL03SDAnd (4) setting standard deviation of elevation difference of corresponding points in all transformation states.
In order to reduce the influence of the ATL03 elevation anomaly data, ATL03 elevation anomaly data points are repeatedly eliminated under different transformation states in the example. Specifically, in different transformation states, photons with abnormal elevation in ATL03 in the state are removed according to formula (4), and then the standard deviation of the corresponding point elevation difference in the transformation state is recalculated according to formula (5).
Figure BDA0003380834250000092
Wherein Hu_ATL03(i, j) high-quality laser height measurement data with abnormal elevation removed in the jth transformation state, Hut_ICESAT2High-quality laser height measurement data sets with the elevation outliers removed in all the transformation states are acquired;
then there are:
ΔHu＝{Δhu(i,j)/Δhu(i,j)＝Hu_ATL03(i,j)-HDSM(i,j),i∈1,2…n,j∈1,2…m}； (5)
wherein, Δ HuFor the recalculated set of elevation differences, Δ h, for all transform statesu(i, j) is the recomputed height difference for the j-th transformation. And finally, selecting the rotation translation amount with the minimum standard deviation as the alpha beta gamma delta x delta y parameter according to the formula (6) and the formula (1). At this time, the height difference average value when the standard deviation of the height difference is minimum is calculated as the translation parameter Δ z in the z-axis direction to obtain the complete 6 transformation parameters. And taking the DSM as source data, and performing inverse transformation according to the 6 transformation parameters so as to improve the elevation precision of the DSM. Wherein, the calculation formula of the standard deviation of the minimum height difference is as follows:
Figure BDA0003380834250000101
wherein,. DELTA.husd(j) The standard deviation of the height difference recalculated under the same transformation state; Δ husd_minThe standard deviation of the smallest elevation difference for all the transition states.
The method is based on statistics, and a target function is established by taking the minimum standard deviation of the height difference as a criterion; compared with the method for iteratively calculating the registration parameters by the least square principle, the method for calculating the registration parameters has the advantage that the defect of a local optimal solution is avoided. Meanwhile, compared with the common ICP, the method is low in complexity and easy to realize, and can resist the influence of a large number of abnormal points while effectively registering the two point clouds.
From 5 pairs of stereo images in the study area, 5 DSMs were extracted by SGM matching method, as shown in fig. 8, a, b, c, d, e are respectively: WV2_ DSM, spatial resolution 1.0m, ZY3_ DSM _ N32E91, spatial resolution 5.0m, GF7_ DSM, spatial resolution 1.5m, ZY3_ DSM _ N39E115, spatial resolution 5.0m, SV _ DSM, spatial resolution 1.5 m. A WGS84 model is selected as a coordinate system of DSM, and an elevation datum plane is the height of a WGS84 ellipsoid and is consistent with a space coordinate system of ATL 03. In order to fully and specifically evaluate the effect of the registration method used in the method, the optimized result of the DEM is compared and analyzed with the ICP registration result from plane, vertical direction and point to plane.
In the two-dimensional plane direction, the optimized DSM center movement vector in the plane direction is registered by comparing an ICP algorithm for analysis with the method of the present example, so that the effect of the DSM registration algorithm based on terrain similarity is evaluated. Table 2 shows the translation amount of the DSM in the right east and north directions after the registration optimization of the 5DSM by applying the ICP algorithm and the algorithm of this example, and the difference of the translation amount is shown. It can be seen that after two kinds of registration algorithms are optimized for WV2_ DSM, SV _ DSM, and GF7_ DSM, the difference in the amount of translation of the DSM center in the east and north directions reaches the sub-meter level, which is smaller than the spatial resolution of DSM, and the difference in the amount of translation of the DSM center in the east and north directions is larger than 1m for ZY3_ DSM _ N32E91 and ZY3_ DSM _ N39E115, but is also smaller than the spatial resolution of DSM. It can be seen that, when the allowable error is less than one pixel, the registration algorithm of the present example has an effect equivalent to that of the ICP point cloud registration algorithm in the plane direction, and the smaller the DSM spatial resolution is, the closer the registration effect of the present example is to that of the ICP algorithm. Fig. 9 shows the spatial distribution of the translation amount of the DSM center plane after the registration optimization of the algorithm and the ICP algorithm is applied, and the difference of the translation vector of the DSM center plane after the optimization of the DSM by the two algorithms can be more visually seen compared with the form of a table.
Table 2 DSM center plane translation amount units after registration optimization: m is
Figure BDA0003380834250000111
In the vertical direction, the effect of the terrain similarity (BOTS) DSM registration algorithm is evaluated by adopting a cross validation and comparison analysis method. The specific scheme of the cross validation is as follows: and selecting laser points uniformly distributed on the whole DSM from the finally screened ATL03 data as elevation check points for counting the elevation residual error of the DSM and evaluating elevation precision, wherein the rest points are used for optimizing the geometric precision of the DSM. Fig. 12 shows the spatial distribution of high range checkpoints in various DSMs. In fig. 10, a, b, c, d and E respectively count percentiles and mean values of absolute values of elevation residuals before and after optimization by two registration algorithms, namely, WV2_ DSM, ZY3_ DSM _ N32E91, GF7_ DSM, ZY3_ DSM _ N39E115 and SV _ DSM. Wherein, elevation residual errors of all DSMs are greatly improved after optimization by two registration algorithms; for DSMs optimized by applying two mating algorithms, the residual variation is small between 0% and 75% of DSM elevation residual errors, which indicates that the elevation residual error distribution is concentrated, but when the DSM elevation residual errors are between 75% and 100% percentiles, the residual variation suddenly changes greatly, which indicates that the elevation residual error distribution is discrete, so that the elevation residual error points between 75% and 100% percentiles are considered to belong to abnormal data. In addition, at each percentile, the elevation residual absolute value and the elevation residual absolute average value of each DSM after optimization by two hybridization algorithms are approximately consistent, which shows that the DSM registration algorithm Based On Terrain Similarity (BOTS) in the vertical direction can achieve the effect of ICP algorithm.
According to the analysis result of fig. 10, after the abnormal elevation residual data between 75% and 100% percentiles are removed, the Root Mean Square Error (RMSE) of the elevation residual in each DSM before and after optimization by applying two registration algorithms is counted in table 3, the elevation accuracy of each DSM is evaluated by using the Root Mean Square Error (RMSE) as an index, and the effect of DSM registration algorithm Based On Terrain Similarity (BOTS) in this example is illustrated by comparing RMSE. Table 3 shows that after the exception checkpoint is eliminated, the relative size of each unregistered optimized DSM elevation residual RMSE value matches the actual value; after optimization through an ICP algorithm and a BOTS algorithm, the elevation residual RMSE values of the WV2_ DEM and the SV _ DEM reach a sub-meter level, theoretically, the elevation residual RMSE of the GF7_ DSM can also reach the sub-meter level, but because dense vegetation covers the GF7_ DSM area, the elevation quality of the GF7_ DSM and the ATL03 is further lost, and the elevation residual RMSE of the GF7_ DSM is larger than 1 m; for ZY3_ DSM _ N32E91 and ZY3_ DSM _ N39E115, it is believed that the reason for the small spatial resolution results in elevation residual RMSE values greater than 1 m.
Table 3 DEM elevation residual RMSE statistical units: m is
Figure BDA0003380834250000121
Fig. 11 visually compares RMSE values of DSM elevation residuals before and after optimization of each DSM using two registration algorithms, and in combination with the statistical analysis of table 3, shows that the elevation residual RMSE of DSM after DSM registration algorithm Based On Terrain Similarity (BOTS) in the vertical direction shows significant advantages over the elevation residual RMSE of DSM that is not optimized for registration, and the effect is equivalent to that of ICP algorithm. The magnitude of the optimization differs only because the imaging conditions of each stereo pair differ.
Fig. 12 a, b, c, d, E respectively show spatial distribution of absolute values of elevation residuals after WV2_ DSM, ZY3_ DSM _ N32E91, GF7_ DSM, ZY3_ DSM _ N39E115, and SV _ DSM are optimized by a terrain similarity (BOTS) -based terrain three-dimensional point cloud registration algorithm. It can be seen that the check points are distributed on each DSM basically and uniformly, and the reliability of DSM elevation precision analysis is ensured. The DSM elevation optimization result is greatly influenced by terrain, ground surface coverage type and DSM spatial resolution. The points with large elevation residual errors in each DSM are mostly distributed in mountainous areas with steep terrain and complex terrain, and the elevation residual errors in the smooth areas are small.
In conclusion, due to the development of the position and orientation measurement technology of the earth orbit space sensor, the accuracy of direct earth positioning of most high-resolution satellite remote sensing images is better than 10m at present, but the requirements of many high-accuracy earth measurement applications and geographic environment monitoring researches cannot be met. In satellite photogrammetry, a small number of control points are usually added in the block adjustment to improve the positioning accuracy of a satellite remote sensing image, but in the work flow, the positions of the control points in the image need to be accurately identified, and the control points cannot be measured on the spot in a complicated terrain area. Accordingly, the present invention introduces a terrain similarity (BOTS) based DSM and ICESat-2 laser altimetry data registration algorithm to optimize the geometric accuracy of the satellite DSM. Experiments show that in the two-dimensional plane direction, the difference between the registration accuracy of the BOTS algorithm and the ICP algorithm is smaller than one pixel, and the higher the DSM resolution is, the smaller the difference is, and the registration accuracy identical to that of the ICP method can be achieved within an error allowable range; and the registration accuracy of the BOTS algorithm and the ICP algorithm is nearly consistent in the vertical direction. From the aspect of algorithm complexity, the BOTS algorithm is simpler and easier to implement than the ICP algorithm. The BOTS algorithm obtains the registration parameters by traversing all possible coordinates, so that the limitation of the local optimal solution of the ICP algorithm is avoided. In experiments, ICP registration was found to fail when there were a large number of outliers in the ATL03 data or the ATL03 was a large distance from the initial spatial pose of the DSM, showing that the registration method of this example is more robust. The optimized DSM elevation RMSE (root mean square error) fused with the ICESat-2 can reach a sub-meter level, the elevation precision is improved by 73-92%, wherein the elevation RMSE after optimization of the Worldview-2 DSM is 0.71 m.

Claims (10)

1. A method for improving elevation accuracy of an optical satellite stereoscopic image DSM is characterized by comprising the following steps: creating a DSM (digital surface model) by using a satellite stereoscopic image to acquire satellite laser height measurement data of the same area; and after the space coordinate transformation is carried out on the two, the registration is carried out, and the registration state with the minimum standard deviation of the elevation difference of the two is taken as the registration result, so that the elevation precision of the DSM is improved.
2. The method as claimed in claim 1, wherein when creating DSM by using satellite stereogram, the image geometric model is RFM model, the output resolution is no more than 2 times the image resolution, and is stored in regular grid form.
3. Method for increasing the elevation accuracy of optical satellite stereoscopic imagery DSM according to claim 1 or 2, comprising the steps of: the method comprises the steps of obtaining satellite laser height measurement data of the same area, preprocessing the satellite laser height measurement data to obtain high-quality laser height measurement data, and then performing space coordinate transformation; the pretreatment comprises the following steps:
weak energy laser beam data in the satellite laser height measurement data are removed, and medium and high-confidence photons when the earth surface coverage type is the land are selected from the rest data; the medium and high confidence photons are linearly gathered and distributed along the laser track, the medium and high confidence photons are distributed on a line segment perpendicular to the direction of the laser track, and the average elevation value of the photons in the same line segment is taken as the final elevation of the line segment; and removing photons with abnormal elevation.
4. The method for improving the elevation accuracy of the optical satellite stereoscopic image DSM according to claim 3, wherein the step of eliminating the photons with abnormal elevation comprises the following steps:
and setting a threshold value of the height difference between the photons and the DSM, and removing the photons exceeding the threshold value as the photons with abnormal height.
5. The method according to claim 4, wherein the method for improving the elevation accuracy of the optical satellite stereoscopic image DSM comprises the following steps of performing spatial coordinate transformation on the optical satellite stereoscopic image DSM and the optical satellite stereoscopic image DSM, performing registration, and taking a registration state with the minimum standard deviation of the elevation difference as a registration result to improve the elevation accuracy of the DSM:
after space coordinate transformation is carried out on DSM and high-quality laser height measurement data, the maximum translation rotation amount is set, the corresponding point relation is established, the high-quality laser height measurement data is used as source data, DSM is used as reference data, and space rotation transformation is carried out around an X, Y, Z axis; translation transformation is carried out in an X-O-Y plane, and the standard deviation of the height difference of the corresponding point under different transformation states is calculated;
selecting the rotation translation amount with the minimum standard deviation of the height difference as a transformation parameter;
and taking the DSM as source data, and performing inverse transformation according to the transformation parameters so as to improve the elevation precision of the DSM.
6. The method for improving the elevation accuracy of the optical satellite stereoscopic image DSM according to claim 5, further comprising the steps of: and under different transformation states, rejecting photons with abnormal elevation in the high-quality laser height measurement data under the state, and recalculating the standard deviation of the height difference of the corresponding point under the transformation state.
7. The method according to claim 6, wherein the standard deviation of the elevation difference of the corresponding point under different transformation states is calculated as follows:
ΔH＝{Δh(i,j)/Δh(i,j)＝HATL(i,j)-HDSM(i,j),i∈1,2…n,j∈1,2…m}；
wherein HATL(i,j)、HDSM(i, j) respectively indicating the elevations of corresponding points formed by laser points of high-quality laser height measurement data and DSMs, wherein delta H (i, j) is the elevation difference of the corresponding points after each transformation, and delta H is an elevation difference set of the corresponding points in all transformation states; i is the number of corresponding points, and j is the number of times of executing space transformation on the high-quality laser height measurement data;
Figure FDA0003380834240000021
wherein,. DELTA.hsd(j) Standard deviation of elevation difference, delta H, of corresponding point under jth spatial transformation for high-quality laser altimetry dataSDAnd (4) setting standard deviation of elevation difference of corresponding points in all transformation states.
8. The method according to claim 7, wherein the transformation parameters are calculated as follows:
Figure FDA0003380834240000022
wherein T is a translation matrix; r is a rotation matrix; alpha, beta, gamma, delta x, delta y and delta z are respectively the rotation translation amount of the x axis, the y axis and the z axis; the translation parameter Δ z in the z-axis direction is an elevation difference average value when the standard deviation of the elevation difference is minimum.
9. The method for improving the elevation accuracy of the optical satellite stereoscopic image DSM according to claim 8, wherein the calculation formula of removing the photons with elevation anomaly is as follows:
Figure FDA0003380834240000023
wherein Hu_ATL(i, j) high-quality laser height measurement data with abnormal elevation removed in the jth transformation state, Hut_ICESAT2High-quality laser height measurement data sets with the elevation outliers removed in all the transformation states are acquired;
then there are:
ΔHu＝{Δhu(i，j)/Δhu(i，j)＝Hu_ATL(i，j)-HDSM(i，j)，i∈1，2…n，j∈1，2…m}；
wherein, Δ HuFor the recalculated set of elevation differences, Δ h, for all transform statesu(i, j) is the recomputed height difference for the j-th transformation.
10. The method according to claim 9, wherein the calculation formula of the standard deviation of the minimum elevation difference is as follows:
Figure FDA0003380834240000024
wherein,. DELTA.husd(j) The standard deviation of the height difference recalculated under the same transformation state; Δ husd_minThe standard deviation of the smallest elevation difference for all the transition states.
CN202111433232.2A 2021-11-29 2021-11-29 Method for improving elevation precision of optical satellite stereoscopic image DSM Active CN114689015B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111433232.2A CN114689015B (en) 2021-11-29 2021-11-29 Method for improving elevation precision of optical satellite stereoscopic image DSM

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111433232.2A CN114689015B (en) 2021-11-29 2021-11-29 Method for improving elevation precision of optical satellite stereoscopic image DSM

Publications (2)

Publication Number Publication Date
CN114689015A true CN114689015A (en) 2022-07-01
CN114689015B CN114689015B (en) 2023-01-17

Family

ID=82135584

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111433232.2A Active CN114689015B (en) 2021-11-29 2021-11-29 Method for improving elevation precision of optical satellite stereoscopic image DSM

Country Status (1)

Country Link
CN (1) CN114689015B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114859374A (en) * 2022-07-11 2022-08-05 中国铁路设计集团有限公司 Newly-built railway cross measurement method based on unmanned aerial vehicle laser point cloud and image fusion
CN115808675A (en) * 2023-01-17 2023-03-17 湖南迈克森伟电子科技有限公司 Laser ranging error compensation method
CN116188497A (en) * 2023-04-27 2023-05-30 成都国星宇航科技股份有限公司 Method, device, equipment and storage medium for optimizing generation of DSM (digital image model) of stereo remote sensing image pair

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103093459A (en) * 2013-01-06 2013-05-08 中国人民解放军信息工程大学 Assisting image matching method by means of airborne lidar point cloud data
CN106960174A (en) * 2017-02-06 2017-07-18 中国测绘科学研究院 High score image laser radar vertical control point is extracted and its assisted location method
CN107167786A (en) * 2017-06-05 2017-09-15 中国测绘科学研究院 Laser satellite surveys high data assisted extraction vertical control point method
CN110487241A (en) * 2019-08-15 2019-11-22 中国测绘科学研究院 Laser satellite surveys high extraction building area vertical control point method
CN111126148A (en) * 2019-11-25 2020-05-08 长光卫星技术有限公司 DSM (digital communication system) generation method based on video satellite images
CN111156960A (en) * 2019-12-28 2020-05-15 同济大学 Satellite laser elevation control point screening method suitable for unstable ground surface area
CN112381940A (en) * 2020-11-27 2021-02-19 广东电网有限责任公司肇庆供电局 Processing method and device for generating digital elevation model from point cloud data and terminal equipment
CN112489212A (en) * 2020-12-07 2021-03-12 武汉大学 Intelligent three-dimensional mapping method for building based on multi-source remote sensing data
KR102237097B1 (en) * 2021-01-12 2021-04-08 헬리오센 주식회사 Transformation system of DEM with aircraft photographing image from DEM by using AI
CN112985358A (en) * 2021-02-19 2021-06-18 武汉大学 ICESat-2/ATLAS global elevation control point extraction method and system
CN113379648A (en) * 2021-07-09 2021-09-10 自然资源部国土卫星遥感应用中心 High-resolution seven-and-resource three-dimensional image joint adjustment method
CN113514829A (en) * 2021-07-12 2021-10-19 自然资源部国土卫星遥感应用中心 InSAR-oriented initial DSM block adjustment method
CN113538501A (en) * 2021-08-24 2021-10-22 荆门汇易佳信息科技有限公司 Low-altitude image DSM generation building edge refinement method

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103093459A (en) * 2013-01-06 2013-05-08 中国人民解放军信息工程大学 Assisting image matching method by means of airborne lidar point cloud data
CN106960174A (en) * 2017-02-06 2017-07-18 中国测绘科学研究院 High score image laser radar vertical control point is extracted and its assisted location method
CN107167786A (en) * 2017-06-05 2017-09-15 中国测绘科学研究院 Laser satellite surveys high data assisted extraction vertical control point method
CN110487241A (en) * 2019-08-15 2019-11-22 中国测绘科学研究院 Laser satellite surveys high extraction building area vertical control point method
CN111126148A (en) * 2019-11-25 2020-05-08 长光卫星技术有限公司 DSM (digital communication system) generation method based on video satellite images
CN111156960A (en) * 2019-12-28 2020-05-15 同济大学 Satellite laser elevation control point screening method suitable for unstable ground surface area
CN112381940A (en) * 2020-11-27 2021-02-19 广东电网有限责任公司肇庆供电局 Processing method and device for generating digital elevation model from point cloud data and terminal equipment
CN112489212A (en) * 2020-12-07 2021-03-12 武汉大学 Intelligent three-dimensional mapping method for building based on multi-source remote sensing data
KR102237097B1 (en) * 2021-01-12 2021-04-08 헬리오센 주식회사 Transformation system of DEM with aircraft photographing image from DEM by using AI
CN112985358A (en) * 2021-02-19 2021-06-18 武汉大学 ICESat-2/ATLAS global elevation control point extraction method and system
CN113379648A (en) * 2021-07-09 2021-09-10 自然资源部国土卫星遥感应用中心 High-resolution seven-and-resource three-dimensional image joint adjustment method
CN113514829A (en) * 2021-07-12 2021-10-19 自然资源部国土卫星遥感应用中心 InSAR-oriented initial DSM block adjustment method
CN113538501A (en) * 2021-08-24 2021-10-22 荆门汇易佳信息科技有限公司 Low-altitude image DSM generation building edge refinement method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
叶江: "面向青藏高原矿集区三维场景的高分辨率卫星影像精处理方法", 《中国博士学位论文全文数据库》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114859374A (en) * 2022-07-11 2022-08-05 中国铁路设计集团有限公司 Newly-built railway cross measurement method based on unmanned aerial vehicle laser point cloud and image fusion
CN114859374B (en) * 2022-07-11 2022-09-09 中国铁路设计集团有限公司 Newly-built railway cross measurement method based on unmanned aerial vehicle laser point cloud and image fusion
CN115808675A (en) * 2023-01-17 2023-03-17 湖南迈克森伟电子科技有限公司 Laser ranging error compensation method
CN116188497A (en) * 2023-04-27 2023-05-30 成都国星宇航科技股份有限公司 Method, device, equipment and storage medium for optimizing generation of DSM (digital image model) of stereo remote sensing image pair

Also Published As

Publication number Publication date
CN114689015B (en) 2023-01-17

Similar Documents

Publication Publication Date Title
CN110058237B (en) InSAR point cloud fusion and three-dimensional deformation monitoring method for high-resolution SAR image
CN114689015B (en) Method for improving elevation precision of optical satellite stereoscopic image DSM
US20200103530A1 (en) Method for extracting elevation control point with assistance of satellite laser altimetry data
CN104931022B (en) Satellite image stereoblock adjustment method based on spaceborne laser altimeter system data
CN109100719B (en) Terrain map joint mapping method based on satellite-borne SAR (synthetic aperture radar) image and optical image
CN110889899B (en) Digital earth surface model generation method and device
Aguilar et al. Geometric accuracy assessment of QuickBird basic imagery using different operational approaches
Eltner et al. Integrated processing of high resolution topographic data for soil erosion assessment considering data acquisition schemes and surface properties
Osama et al. A digital terrain modeling method in urban areas by the ICESat-2 (Generating precise terrain surface profiles from photon-counting technology)
CN113238228B (en) Three-dimensional earth surface deformation obtaining method, system and device based on level constraint
Feng et al. An improved geometric calibration model for spaceborne SAR systems with a case study of large-scale Gaofen-3 images
Qiao et al. Assessment of geo-positioning capability of high resolution satellite imagery for densely populated high buildings in metropolitan areas
Recla et al. From Relative to Absolute Heights in SAR-based Single-Image Height Prediction
Saldaña et al. DSM Extraction and evaluation from geoeye-1 stereo imagery
CN116977580A (en) Method for manufacturing mountain area large scale DEM based on airborne LiDAR
Yastikli et al. Quantitative assessment of remotely sensed global surface models using various land classes produced from Landsat data in Istanbul
CN113720351A (en) Joint area network adjustment method for satellite-borne laser height measurement data and remote sensing stereo image
Murthy et al. Analysis of DEM generated using Cartosat-1 stereo data over Mausanne Les Alpiles–Cartosat scientific appraisal programme (CSAP TS–5)
Pradhan et al. Elevation modeling using radargrammetry: Case study from Malaysia
Zhao et al. Comparison and analysis of accuracy of elevation extraction based on the ZY-3 01 and 02 satellites stereoscopic images
CN118519132B (en) High-precision satellite-borne single-photon laser calibration method based on corner reflector
Zhang et al. Satellite remote sensing image stereoscopic positioning accuracy promotion based on joint block adjustment with ICESat-2 laser altimetry data
CN113450348B (en) Soil erosion quantitative estimation method based on high-resolution stereopair image
CN113900069B (en) Perpendicular deviation calculation method and system based on interference imaging altimeter
Clay et al. Assessing the Accuracy of Georeferenced Point Clouds from Uas Imagery

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant